Safety Trustworthiness: Protect people, optimize production, secure machines

Related Vendors

FAULHABER Drive Systems

Wuxi InfiMotion Technology Co., Ltd.

Dynamic safety aims to respond flexibly to hazardous situations in order to maintain maximum production speed. The companies Pilz, TÜV Süd and B&R Automation are working together on the safety use case of the SmartFactory-KL.

Ingo Herbst is Head of Communication and Press Spokesman of SmartFactory-KL.

Specifically, the use case consists of two scenarios: a machine has a potential danger point and there is a defective hazardous material in the production process.

Case Hazard Point
In this case, Operational Safety Intelligence should ensure that there is no danger to employees. A laser scanner and a radar are combined, which constantly monitor the area in front of the danger point.

With an AI layer, as a complement to classic safety, it can be analyzed how a person moves in the danger zone. Is he approaching, or is he passing by? Depending on the distance of a person to the danger zone, the system reacts flexibly. It starts with the constant reduction of speed up to a total standstill. Thus, the machine is only completely shut down in exceptional cases, ultimately resulting in an overall increase in production. It can also distinguish between an AMR/AGV and a human being, with the system not slowing down production when a vehicle approaches.

Case Hazardous Material
A hazardous material is transported in production. In this particular case, it is a lithium-ion battery. It moves as quickly as possible in the production process with adjusted speed. If it is recognised as defective during quality control, it is immediately removed from production and handed over to an employee by the shortest route.

The SmartFactory-KL works with the term Trustworthiness in its Use Cases. Behind this is the idea of pointing out the interaction of OT and IT. Because theoretically, errors from machines, people, or data can lead to information that is illogical and thus untrustworthy. An AI recognizes this anomaly and points it out. This prevents, for example, a hazardous material from suddenly being classified as a harmless object or a person approaching is misinterpreted as an AGV.

Focus Hazardous Material: Safety meets Security in Industry 4.0 Ecosystems

The production island KUBA is a multi-agent system that uses digital twins and the Asset Administration Shell (AAS) to create a digital representation for each physical element. This digital representation allows for individual control and monitoring of the production and transport process of hazardous goods.

The first step in the use case is the introduction of an empty truck trailer into the KUBA production island system. The trailer is initially classified as safe, allowing normal productivity without special restrictions. As soon as the empty trailer is filled with lithium-ion batteries in the assembly module, the system status changes: the Product-AAS now recognizes a potential hazard from the batteries, which requires appropriate handling of the trailer. All other shuttles are not affected.

To make the demonstration feasible at a trade fair, we rely on an optical inspection of the battery packs to simulate errors, rather than on real measurements. In real production, these errors can be detected by continuously monitoring the condition of the batteries. If an error is visually identified, the affected battery pack is classified as actively dangerous. The Product-AAS is updated to reflect the changed hazard class and immediately initiates safety protocols, such as slowing down and diverting the specific shuttle for error analysis and correction by an employee. After the correction and re-inspection, the product is re-categorized so that the individual speed can be increased again and the production process can be completed.

Safe handling requires security.

To ensure the safe handling of dangerous goods, a specialized architecture based on digital twins is essential, but just as critical is ensuring the accuracy and integrity of the data. Protection through cybersecurity methods is essential, because without this continuous monitoring, security can be compromised. Practical examples of potential threats include not only cyber attacks that can compromise the network infrastructure and cyber-physical production systems (CPPS) from the outside, but also internal processes such as maintenance work or the introduction of unknown devices into the system by maintenance engineers - whether unintentional or deliberate.

Particularly in the area of transport of dangerous goods, where system-critical parameters form the foundation of safety mechanisms, reliable cybersecurity is of utmost importance. A scenario where an error in the lithium-ion batteries is detected during operation must not lead to the failure of the ejection mechanism due to parameter errors. Such a failure could undermine safety and, in the worst case scenario, lead to serious consequences such as a battery fire - with potentially catastrophic consequences for the production site.
To handle the complexity of modern production environments and the associated risk potential, a continuously active anomaly detection is integrated into our system. It forms a central component of our safety architecture and continuously monitors the system parameters and the connection of new and potentially unknown devices to the network. In case of irregularities, automated warning messages are generated to inform the operator immediately.

People play a key role in our safety concept: As the last instance, they make decisions and initiate the necessary measures based on the information provided by the anomaly detection system. This process allows for targeted error analysis and helps the operating personnel to identify and understand the causes of any problems that may arise.